AISI Cyber CTF: leaderboard
AISI's 71-task private CTF suite where autonomous agents solve web exploitation, cryptography, reverse engineering, forensics, and privilege-escalation challenges.
Metric: Success Rate (%). Source: arxiv.org. Saturation forecast: Around January 2027. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.3 Codex | 75 | #69 |
| 2 | Claude Opus 4.6 | 72 | #63 |
| 3 | GPT-5.2 Codex | 72 | #95 |
| 4 | GPT-5 | 65 | #98 |
| 5 | Claude Opus 4.5 | 63 | #83 |
| 6 | O3 | 54 | #124 |
| 7 | Claude Sonnet 4.5 | 45 | #142 |
| 8 | Claude Opus 4 | 39 | #163 |
| 9 | O1 | 38 | #183 |
| 10 | Claude 3.7 Sonnet | 34 | #249 |
| 11 | Claude 3.5 Sonnet | 30 | #350 |
| 12 | GPT-4 Turbo | 18 | #404 |
| 13 | Claude 3 Opus | 16 | #571 |
| 14 | GPT-4o | 15 | #343 |
| 15 | GPT-3.5 Turbo | 6 | #865 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=aisi-cyber-ctf · How It Works · Data refreshed daily, snapshot 2026-10-20.